Policy Research Working Paper 10707 Presumptive Tax on Small and Microenterprises with a Gender Lens in Ethiopia Hitomi Komatsu Development Economics A verified reproducibility package for this paper is Development Data Group available at http://reproducibility.worldbank.org, February 2024 click here for direct access. Policy Research Working Paper 10707 Abstract Governments often use simplified business tax systems, three key findings. First, the effective tax rates are higher for such as presumptive tax regimes, to register and tax small businesses in the lowest quartile at 4.3 percent of turnover and microenterprises. Despite concerns about how such compared to 1.5 percent for businesses in the highest quar- regimes could disproportionately affect female-owned and tile, using the most recent survey, resulting in a regressive low-revenue entrepreneurs, there is a lack of empirical system. Second, male-owned businesses tend to operate in analysis examining the tax burden. The presumptive tax sectors without a tax-free threshold and are more likely than in Ethiopia has a complex assessment system, where the female-owned businesses to face higher tax rates. Third, the tax liabilities are determined according to the activity type effective tax rates are high for businesses in food and bev- and turnover (99 activities and 19 turnover bands), and erage services, which is female-dominated, and for transit some activities do not have a tax-free threshold. This paper services, which is male-dominated, due to the lack of a uses two rounds of data in the Ethiopian Socioeconomic tax-free threshold for these sectors. The study finds that an Surveys and the tax code to analyze the equity and gender alternative presumptive tax system with a single tax rate on implications of the presumptive tax on small and micro- turnover and an exemption for all low-revenue businesses enterprises by imputing the effective tax rates. There are would be simpler for tax assessment and more progressive. This paper is a product of the Development Data Group, Development Economics. . It is part of a larger effort by the World Bank to provide open access to its research and make a contribution to development policy discussions around the world. Policy Research Working Papers are also posted on the Web at http://www.worldbank.org/prwp. The author may be contacted at hkomatsu@worldbank.org. A verified reproducibility package for this paper is available at http:// reproducibility.worldbank.org, click here for direct access. RESEA CY LI R CH PO TRANSPARENT ANALYSIS S W R R E O KI P NG PA The Policy Research Working Paper Series disseminates the findings of work in progress to encourage the exchange of ideas about development issues. An objective of the series is to get the findings out quickly, even if the presentations are less than fully polished. The papers carry the names of the authors and should be cited accordingly. The findings, interpretations, and conclusions expressed in this paper are entirely those of the authors. They do not necessarily represent the views of the International Bank for Reconstruction and Development/World Bank and its affiliated organizations, or those of the Executive Directors of the World Bank or the governments they represent. Produced by the Research Support Team Presumptive Tax on Small and Microenterprises with a Gender Lens in Ethiopia 1 Hitomi Komatsua Keywords: Presumptive tax, taxation, gender, Ethiopia JEL: H22, H25, J16 1 This study was supported by the World Bank’s Global Tax Program. I would like to thank Alemayehu Ambel, Silver Namunane, Abhilasha Sahay, Hibret Belete Maemir, Firew Woldeyes, Caren Grown, and Gayatri Koolwal for valuable comments on earlier drafts. Any error or omission is my responsibility alone. a World Bank, Washington, DC, hkomatsu@worldbank.org. 1. Introduction Governments in low- and middle-income countries are increasingly focusing on registering and taxing informal small and microenterprises (Dom and Prichard 2022, Ohnsorge and Yu 2021). Simplified or small business tax regimes that use simple and objective methods for tax assessments are often established for this purpose (Loeprick 2009). A presumptive tax on small and microenterprises is one of the simplified business tax regimes and it uses an indirect method of measuring tax liabilities, which can be assessed on turnover or other indicators (e.g., the number of tables, seats, or floor space) (Loeprick 2009). The rationale for a presumptive tax includes facilitating tax compliance, lowering compliance costs, and supporting formalization of businesses due to the simpler reporting requirements compared to the regular tax system (Mas-Montserrat et al., 2023); enabling revenue authorities to gain valuable information from businesses and to help them transition into the corporate tax system (Engelschak and Loeprick 2016); potentially raising tax revenue given the large informal sector in low- and middle-income countries (Joshi, Prichard and Heady, 2014); and increasing tax morale and tax compliance among formal taxpaying firms (Joshi, Prichard and Heady, 2014). Women tend to be overrepresented in small and microenterprises and thus may be subject to the presumptive tax. In Ethiopia, the presumptive tax is relevant because women predominate in nonfarm enterprises as it is considered inappropriate for women to carry out certain agricultural tasks and there are limited income-earning opportunities elsewhere (Loening, Rijkers, and Söderbom, 2008, Rijkers and Costa 2013). Women’s concentration in some sectors, such as hospitality, personal services, and food processing, may also have gender implications on the tax burden if the tax rates differ by sector. This is the case in Ethiopia where the presumptive tax has a complex assessment system with divergent tax rates by the activity type and turnover - with 99 activities and 19 turnover bands. Yet, an empirical analysis of how the presumptive tax affects low-income and female entrepreneurs has received limited attention because of a lack of data and the low priority given by governments that often generate very little revenue from the presumptive tax (Mas-Montserrat et al., 2023). Despite this, the tax can be burdensome for small and microenterprises (Bucci 2020). Given governments’ keen interest to register informal small businesses and to bring them into the tax net, it is important that the tax burdens on these businesses be examined. This paper examines the gender and equity implications of the presumptive tax on small and microenterprises in Ethiopia by estimating the effective tax rates, disaggregated by the sex of the business owner. The study uses newly available data on household nonfarm enterprises from the 2018/2019 and 2021/2022 Ethiopian Socioeconomic Surveys (ESS) and the tax code to impute tax liabilities. The ESS also collected information on self-reported tax payments by household enterprises. The effective tax rates are 2 calculated by dividing the tax liabilities by the turnover using two different measures of tax liabilities: the self-reported taxes and imputed taxes. 2 Estimating the effective tax rate is relevant because high tax rates could discourage businesses from registering and complying with payments. For example, a survey of informal and formal firms in Ethiopia found high tax rates to be a disincentive for informal firms to formalize and pay taxes (World Bank 2016). Similarly, a study in Pakistan showed that increasing tax rates resulted in formal firms to become informal (Waseem 2018), and in Brazil, lowering tax rates encouraged more firms to formalize (Rocha, Ulyssea, and Rachter 2018). There are three key findings from this study. First, the effective tax rates for the smallest businesses are high because the tax-free threshold is not available for some subsectors. Using the tax code, the effective tax rates for businesses in the lowest turnover quartile are 4.3% of turnover compared to 1.5% for businesses in the highest quartile, using the most recent data, resulting in a regressive system. Second, the effective tax rates in the lowest quartile are higher for male businesses than for female businesses because the former tend to operate in sectors without a tax-free threshold. Third, the effective tax rates for businesses in food and beverage services – a female-dominated sector, and in transit services – a male-dominated sector, are high due to the lack of a tax-free threshold for these sectors. Given the regressivity of the tax code, this paper conducts a policy simulation by estimating the tax burden of a hypothetical presumptive tax regime by exempting all businesses with a revenue of 50,000 birr or less and taxing a fixed rate on turnover for businesses earning above the threshold. This paper finds that a fixed tax rate of 2.5 percent on turnover and exemptions for low revenue firms would be simpler for tax assessment and more progressive. This paper makes a contribution to the literature in three ways. First, the literature studying the tax burden of the presumptive tax is limited because of the lack of disaggregated data and the low priority given by governments as aforementioned. Among the exceptions are empirical studies by: Waiswa et al. (2021), who find some regressivity in the recently reformed presumptive tax in Uganda; Semboja (2015), who shows there is regressivity within some turnover bands in Tanzania; and Engelschalk and Loeprick (2016), who examine the policy design of simplified tax regimes of micro and small businesses in transition economies and conclude that these regimes often do not adequately consider tax fairness. Using a de jure analysis, Dube and Casale (2019) find the presumptive tax in Zimbabwe to be regressive with a particularly high tax burden for hairdressers, who are predominantly women. This paper builds on these studies by examining the equity and gender implications of the tax in Ethiopia. The explicit consideration of the different tax burdens on women and men in the Ethiopian context is important given that the activity- specific tax rates in the tax code and the gendered sectoral distribution of activities are likely to interact. 2 The literature that estimates the effective corporate tax rates uses turnover, gross profits, net profits, or pre-tax income in the denominator (see Bachas et al. 2023 for a review). This paper uses turnover because it is the tax base used to calculate the presumptive tax. 3 Second, there is an emerging body of literature that finds that small firms or traders are required to pay multiple taxes or fees, such as trade association fees, business licenses, storage fees, and market fees in the Democratic Republic of Congo (Paler et al., 2017), Ghana (Anyidoho et al., 2022, Caroll 2011, Prichard and van den Boogaard 2017), Nigeria (Akpan and Cascant-Sempere 2022 and Meagher 2016), Sierra Leone (van den Boogaard 2018, van den Boogaard, Prichard, and Jibao 2019), Uganda (Pimhidzai and Fox 2013), and Zimbabwe (Ligomeka 2019). Many of these papers show evidence of regressivity of small taxes and fees on female traders and entrepreneurs. This paper contributes to this literature by focusing specifically on the presumptive tax on small enterprises with a gender lens. Third, this paper makes a methodological contribution by demonstrating how the gender dimension of a small business tax can be evaluated with publicly available survey data, such as the ESS. It also shows how the effective tax rates and revenue implications can be examined through a policy simulation. Surveys typically do not collect data on taxes paid by small and microenterprises, but the ESS fills this data gap. The enterprise module in the ESS data with detailed information on enterprise activities, revenue, and tax questions allows for this type of granular analysis. This paper is organized as follows: section 2 provides background on the presumptive tax in Ethiopia, section 3 describes the data and the descriptive statistics on gender and nonfarm enterprises, and section 4 discusses the empirical methodology. The results on the effective tax rates are discussed in section 5, the policy simulation of a hypothetical presumptive tax is presented in section 6, and section 7 provides the conclusion. 2. The presumptive tax on small and microenterprises: The Ethiopia context Businesses in Ethiopia are classified into three groups for tax administration purposes, namely Category A, B, and C taxpayers. This paper’s focus is on Category C taxpayers, which are unincorporated small and microenterprises, with an annual turnover of less than 500,000 birr (approximately USD 9,040 3) (Harris and Seid 2021). 4 Category C taxpayers pay the presumptive tax, which is calculated using the gross annual turnover, and are not required to keep a book of accounts (Proclamation No. 410/2017). 5 The presumptive tax rates differ by activity and revenue, consisting of lump sum tax liabilities for 99 business sectors with 19 turnover bands (with a total of 1,881 activity-bands) (see Council of Ministers Regulation No. 410/2017 3 1 USD=55.31 Ethiopian birr (exchange rate on October 6, 2023). 4 In contrast, Category A taxpayers are all incorporated businesses (regardless of turnover) and unincorporated businesses with a turnover of more than 1 million birr. Category B taxpayers are unincorporated businesses with a turnover between 500,000 and 1 million birr (Harris and Seid 2021, Mascagni and Molla 2018). 5 There is also an indicator system for operating public buses, freight transport, and flour mills (Council of Ministers Regulation No. 410/2017). 4 and Appendix 1 for the presumptive tax schedule for selected activities). The presumptive tax also includes liberal professionals (such as accountants, auditors, attorneys) provided that their annual earnings are below 500,000 birr. In contrast, in many other countries, the presumptive tax excludes liberal professionals because they are deemed to be able to comply with the accounting requirements of the standard tax regimes (Mas-Montserrat et al., 2023). Some subsistence-level businesses (earning less than 50,000 birr) in certain sectors are subject to the presumptive tax. A tax exemption threshold exists for businesses with a revenue of less than 50,000 birr per year (approx. USD 904) in activities, such as trading in fruit and vegetables, cereal, edible oils, textiles, animal products, and selected services (Council of Ministers Regulation No. 410/2017). However, the tax- free threshold does not exist for food and beverage services (for example, bars, cafes, or hotels), construction, transit services, and trading in specific products (for example, flour, chat, and metals) (Council of Ministers Regulation No. 410/2017). Businesses in the latter group are therefore subject to tax from the moment they earn income. Regional governments are responsible for administration of the presumptive tax (Dixon, Freedman and Yesegat 2019). The revenue generated from small and microbusinesses comprises a small proportion of the regional governments’ tax revenues. For example, in Oromia, it accounts for less than 10 percent of the total tax revenues in 2016/2017 and 2017/2018 (Dixon, Freedman, and Yesegat 2019). Businesses generally do not calculate presumptive tax liabilities themselves, rather, the government sends out revenue assessment teams at the woreda (district) level to estimate the firms’ average daily revenue and extrapolate their annual turnover using an assumed number of working days (Workneh et al., 2019). The door-to-door collection and extrapolation of annual turnover have at times caused taxpayer dissatisfaction (Workneh et al., 2019). Taxpayers deemed the annual turnover estimation to be too high because they did not account for the business seasonality, and therefore overestimated their tax liabilities (Dixon, Freedman, and Yesegat 2019).6 A survey conducted soon after the 2017 reassessment of average daily revenue and presumptive taxes in Addis Ababa found that the assessment process lacked in objectivity, and applied inconsistent methods for estimating the incomes of similar enterprises (Workneh et al., 2019). There was also a lack of understanding of the assessment methods on the part of taxpayers and assessors because the tax schedule, which disaggregates businesses into 1,881 different tax activity-bands, is overly- complicated and subject to different interpretation (Getachew 2019, Workneh et al., 2019). In addition to the presumptive tax, businesses are also subject to the turnover tax, which is required for firms earning below the VAT threshold of 1 million birr (Harris and Seid 2021). The turnover tax rate is 2% of annual turnover for locally sold goods and services provided by contractors, grain mills, combine 6 If revenues are estimated on a longer interval, such as on a monthly basis instead of daily revenue, it could account for seasonality. 5 harvesters, and 10% for other services (Harris and Seid 2021). The lack of an exemption threshold for the turnover tax and the high tax rate for most services make it particularly burdensome for microbusinesses, resulting in taxpayer dissatisfaction (Dixon, Freedman, and Yesegat 2019). In other countries, it is common to introduce a single small business tax or a presumptive tax regime, which replaces the VAT and other income taxes (Engelschalk and Loeprick 2016, IMF 2015). 7 In contrast, in Ethiopia, small and microbusinesses are subject to two taxes – the presumptive tax and the turnover tax - with differing thresholds. 8 3. Data and methodology 3.1 Data The analysis uses two rounds of the Ethiopia Socioeconomic Survey (ESS) 2018/2019 and 2021/2022, a nationally representative panel survey, consisting of 6,770 households in 2018/2019 and 4,959 households in 2021/2022, although the Tigray region was excluded from the ESS 2021/2022. 9 The nonfarm enterprises in the sample are representative of those operated by the households. Period 1 is defined to be the ESS 2018/2019 and period 2 is the ESS 2021/2022. The panel survey includes a nonfarm enterprise module with questions on whether anyone and who in the household owns a nonfarm enterprise, the enterprise activity, the average monthly revenue, and whether the business is registered. It also includes a question on whether the enterprise paid taxes on income as well as the amount paid. The focus on Category C enterprises (with an annual turnover of less than 500,000 birr) resulted in 109 firms being excluded because their revenue exceeded the cut-off. The annual turnover is estimated by multiplying the self-reported average monthly revenue by the number of months of operation in a year. The resulting sample consists of 1,523 nonfarm household enterprises in 2018/2019 and 1,374 enterprises in 2021/2022, which are disaggregated by sex of the owners. Female businesses are defined as those owned exclusively by women, while male businesses are those owned exclusively by men or jointly by women and men. Jointly-owned firms are classified as male firms because the presence of a male owner allows for easier entry into male-dominated sectors, which are more profitable (Alibhai et al., 2017). The estimates in the paper use household weights. 3.2 Descriptive statistics: Gender and nonfarm enterprises 7 See Marchese (2021) for examples of these cases from Latin America. 8 The presumptive tax threshold at 500,000 birr is half the size of the VAT and turnover tax threshold at 1 million birr. 9 There were other ESS surveys that were conducted prior to the current presumptive tax schedule took effect in 2017 and are therefore excluded from the analysis. 6 Nonfarm activity is often the only way for women to generate income and it comprises a significant share of their income in Ethiopia (Fuje 2017, World Bank 2009). This is because it is considered inappropriate for women to engage in certain agricultural activities (such as using oxen and plowing land), which discourages single or widowed women from becoming independent farmers (Bezabih et al., 2016, World Bank 2009). About a quarter of households own a nonfarm enterprise but there is heterogeneity by household type, as shown in Table 1. Households are classified by gender categories where female-only households are defined as those with female adults only whereas male or dual adult households are households with male adults or both male and female adults. In Appendix 2, households are also disaggregated by the sex of the household head. 10 Results show that a higher share of female only households compared to male or dual adult households owned nonfarm enterprises in period 1, but the reverse is observed in period 2. The results are largely consistent when households are disaggregated by sex of the household head (see Appendix 2). In Table 2, when household characteristics are controlled for in a linear probability model, female only households are more likely than male or dual adult households to own an enterprise (which is significant in period 1 and marginally significant at 10% in period 2). In period 1, female-headed households are also more likely than their male counterparts to own an enterprise, but the difference is not significant in period 2. There is also heterogeneity by poverty status: the probability of business ownership increases with household consumption for male or dual adult households, but the probability does not increase as steeply for female only households. Appendix 3 shows that smaller enterprises - measured by their turnover - tend to be owned by poorer households and poor female-only households. The question of whether low revenue enterprises face a disproportionate tax burden therefore is important from an equity and gender perspective. Table 1: Percentage of households with a non-farm enterprise (NFE), by household type Period 1 Period 2 (2018/2019) (2021/2022) % of households with NFEs 22.9 26.8 Number of households 6,770 4,959 Period 1 Period 2 Male or dual Female only Male or dual adult Female only adult household household household household % of households with NFEs 22.5 25.5 27.2 21.2 Number of households 5,518 1,252 4,337 622 10 This paper primarily classifies households by whether there is a male adult, as the assignment of headship is often arbitrary across countries (see World Bank 2018 for a discussion). The presence of a male adult also results in women having better access to land and agricultural labor (see Komatsu et al. 2022). 7 % of poor households with NFEs 16.0 23.0 19.5 19.9 % of nonpoor households with NFEs 27.1 26.6 32.3 22.0 Author’s calculations using ESS 2018/2019 (period 1) and 2021/2022 (period 2). Notes: Female-only households are households with no male adult present. Poor households are defined as those in the bottom 40% of the adult equivalent expenditure distribution. Significant differences at p < 0.05 are shown in bold. Table 2: Probability that a household owns a nonfarm enterprise (1) (2) (3) (4) VARIABLES Period 1 Period 2 Period 1 Period 2 Female only household = 1 0.542*** 0.392* (0.169) (0.236) Log of household consumption 0.062*** 0.047*** 0.063*** 0.047*** (0.012) (0.015) (0.012) (0.016) Female only household * log of household consumption -0.053*** -0.041* (0.017) (0.023) Household size 0.036*** 0.028*** 0.034*** 0.029*** (0.005) (0.005) (0.005) (0.004) Rural -0.181*** -0.204*** -0.181*** -0.205*** (0.021) (0.027) (0.021) (0.027) Female headed household = 1 0.351*** 0.189 (0.133) (0.187) Female headed household * log of household consumption -0.035** -0.021 (0.013) (0.019) Observations 6,770 4,959 6,770 4,959 R-squared 0.075 0.069 0.074 0.069 Source: Author’s calculations using ESS 2018/2019 and 2021/22. Notes: Coefficients are significant at *** p<0.01, **p<0.05, * p<0.1. Coefficients from a linear probability model using OLS regression are presented. OLS regression includes the number of children disaggregated by age groups, household size, marital status of household head, age of household head, whether the household uses electricity for cooking or lighting, dummy variables for regions, and the religion of the household head, but are not shown. Standard errors in parentheses are clustered at the enumeration area. Female-only households are households with no male adult present. Household consumption is the spatially-adjusted adult equivalent scale household consumption. Of the female entrepreneurs, 40% are divorced, widowed, or unmarried, and approximately 40% live in female-headed households (in Appendix 4). There is some evidence that the composition of enterprise- owning households changed between the survey periods. In period 1, 32% of female entrepreneurs live in households without male adults, but this percentage drops to 9.5% in period 2. The absence of male adults in the households is likely to be disadvantageous for women wanting to engage in more profitable activities. Studies have shown that spousal support or male mentors could play an important role in helping female entrepreneurs enter male-dominated and more profitable sectors (Alibhai et al., 2017, World Bank 2022). 8 There is gendered sector segregation in nonfarm activities and differences in business locations. For example, a majority of female enterprises, compared to only a third of male enterprises, operate in or close to their homes (see Appendix 5). In contrast, male businesses are more likely to be mobile (approximately 20% in Appendix 5). Manufacturing or sale of local alcoholic beverages (e.g., araki, tela) is a low-revenue, labor-intensive activity that women predominate (Woldehanna 2009, World Bank 2009). This activity is distinguished from food and beverage services, which include bars, restaurants, cafes, and other food and beverage-related services. Figure 1 shows that women-owned businesses are overrepresented in local beverage manufacturing or sale in both periods and for period 1, in the food and beverage services also. Male-owned businesses, in contrast, are predominant in transportation services. About 40% of male and female enterprises are involved in trade.11 There was also a shift in the sectoral distribution between the two survey periods, which has implications for the presumptive tax. At the end of this section, there is a discussion on how it affects the presumptive tax liabilities. 11 Household enterprises in Tigray region are excluded from period 2, therefore the distributions of businesses in periods 1 and 2 are not directly comparable. However, the results are largely consistent when the sample is restricted to households observed in both periods 1 and 2 (see Appendix 7). 9 Figure 1: Sector of nonfarm enterprises (%) Period 1, 2018/2019 Trade 45.6 44.1 Manufacture or sale of local beverages 1.3 23.4 Manufacturing clothes 9.1 0.3 Manufacturing food 2.0 6.3 Other manufacturing 8.4 2.5 Food and beverage services and hotels 4.8 14.0 Transportation services 6.5 0.1 Other personal services 13.9 6.6 Construction 1.2 0.1 Others 7.2 2.5 0.0 5.0 10.0 15.0 20.0 25.0 30.0 35.0 40.0 45.0 50.0 Male-owned Female-owned Period 2, 2021/2022 Trade 44.0 43.0 Manufacture or sale of local beverages 2.7 28.6 Manufacturing clothes 1.5 5.4 Manufacturing food 2.2 6.7 Other manufacturing 3.4 7.8 Food and beverage services and hotels 5.4 5.3 Transportation services 0.1 12.1 Other personal services 1.02.7 Construction 0.2 8.6 Others 5.6 13.6 0.0 5.0 10.0 15.0 20.0 25.0 30.0 35.0 40.0 45.0 50.0 Male-owned Female-owned Source: Author’s calculations using ESS 2018/2019 (period 1) and ESS2021/2022 (period 2). Most are microenterprises (defined as having five employees or fewer) and 94% of female businesses are sole proprietors (with no employees) compared to 83% of male businesses (Table 3).12 Female enterprises earn less revenue and are less likely than male enterprises to be registered. Manufacturing or sale of local beverages – a female-dominated activity – generates the lowest median annual sales of all the sectors (Table 12 The government defines microenterprises to be those with no more than five employees and assets valued at or less than 1 million birr (World Bank 2021). 10 4), corroborating previous findings by the World Bank (2009). Even when they operate in the same sectors, female businesses generally earn less than male businesses, as seen in the trade sector. Table 3: Enterprise characteristics, by sex of owner Period 1 (2018/2019) Period 2 (2021/2022) Male- Female- Male- Female- owned owned P-values owned owned P-values Microenterprise (%) 96.0 98.8 0.05 97.1 98.0 0.54 Own account (%) 81.6 94.0 0.00 83.4 93.3 0.00 Registered (%) 28.3 12.0 0.00 24.4 7.6 0.00 Urban (%) 56.8 52.1 0.33 51.3 40.0 0.04 Annual sales Median annual sales ('000 birr) 18.0 6.0 28.8 8.0 Average annual sales ('000 birr) 40.6 17.8 0.00 59.4 26.7 0.00 (63.8) (37.9) (82.5) (56.3) % with sales <=50,000 birr 77.0 92.3 0.00 63.4 88.6 0.00 Taxes Paid market tax or municipality fees (%) 8.2 2.6 0.01 11.0 1.8 0.00 Paid tax on business income (%) 15.6 7.7 0.00 18.4 7.2 0.00 Average tax paid on business income (birr, self-reported) 397.2 132.4 0.00 472.6 148.2 0.00 (1630.5) (685.4) (1532.5) (760.4) Number of enterprises 778 745 711 663 % of enterprises in presumptive tax net 44.1 29.9 0.01 61.0 16.8 0.00 Average presumptive tax (birr, imputed) 641.9 224.7 0.00 1222.0 262.5 0.00 (1825.3) (724.7) (472.6) (148.2) % in sectors with tax-free threshold 68.6 74.6 0.26 63.2 91.9 0.00 Number of enterprises 478 458 498 449 Source: Authors’ calculations using ESS 2018/2019 (period 1) and ESS2021/2022 (period 2). Notes: Numbers in parentheses indicate the standard deviation. Significant differences at p < 0.05 are shown in bold. Firms liable for the presumptive tax are defined as those that are estimated to be in the presumptive tax net based on the government’s tax code. Imputed presumptive taxes are calculated using the estimated annual turnover, business activity, and the presumptive tax schedule in the Council of Ministers Regulation No. 410/2017. There are multiple tax rates within each sector, which requires details on business activity type. Some activities, particularly in the retail and wholesale sector, lack sufficient detail to impute presumptive tax, resulting in some enterprises with missing values. Table 4: Enterprise characteristics, by sex of owner and sector Number of enterprises Period 1 Period 2 Period 1 Period 2 Female- Female- Male- Female- Male- Female- Male-owned owned Male-owned owned owned owned owned owned Median annual sales ('000 birr) by sector Manufacturing beverages na 3.2 na 6.0 113 94 Food and beverage services na 9.6 na 40.0 112 61 Retail trade 19.6 9.6 36.0 10.0 355 392 305 392 Transit services 48.0 na 42.0 na 81 108 Average annual sales ('000 birr) by sector Manufacturing beverages na 7.5 na 12.2 113 94 11 Food and beverage services na 16.6 na 54.4 112 61 Retail trade 44.8 22.8 60.7 35.6 355 392 305 392 Transit services 59.1 na 67.6 na 81 108 % of firms with sales <50,000 birr Manufacturing beverages na 99.1 na 99.5 113 94 Food and beverage services na 92.2 na 62.1 112 61 Retail trade 74.0 88.7 60.1 85.6 355 392 305 392 Transit services 50.4 na 56.5 na 81 108 Source: Author’s calculations using ESS 2018/2019 and ESS2021/2022. Note: Significant differences at p < 0.05 are shown in bold. 3.3 Sector composition and implications for the presumptive tax Male-owned businesses are more likely to pay taxes on incomes (17% versus 7% of female enterprises as shown in Table 3) and larger businesses are more likely to pay taxes than smaller businesses (see Appendix 7). By imputing how much presumptive tax would be paid using the tax code, the results show that a lower share of female enterprises (16.8%) would be liable for the presumptive tax compared to male businesses in period 2 (61% in Table 3). Further, between the two surveys, there was a change in the percentage of businesses estimated to be in the tax net (earning revenue above the tax-free threshold). For female enterprises, the share of businesses in the tax net dropped (from 29.9% in period 1 to 16.8% in period 2) because a higher share operated in sectors with a tax-free threshold in period 2. The drop in the share of female businesses in food and beverage services – a sector without a tax-free threshold – contributes to this result. In contrast, for male enterprises, a larger share is operating in sectors without a tax-free threshold – such as transportation services and construction – and there are more firms with earnings exceeding the exemption threshold (50,000 birr p.a.) in period 2, so the proportion estimated to be in the tax net is 17 percentage points higher (44% in period 1 versus 61% in period 2). Disaggregating by turnover quartiles, Table 5 shows that the shifts in sector distribution between the two surveys result in a lower share of the female businesses and a higher share of male businesses in non- exempt sectors (without a tax-free threshold). In Appendix 6, the exit and entry of businesses (for households observed in both periods) show that, compared to female only households, male or dual adult households are more likely to have reported a new business in period 2 and are less likely to have lost an enterprise, resulting in a net gain of businesses for male or dual households. In contrast, there is no change for female households. By sector, 3.6% of female only households lost a food service enterprise, while 1.7% of male or dual households gained a transit enterprise. During the two surveys, several factors may have contributed to the fall in the share of food and beverage service businesses, including the conflict in Ethiopia, the pandemic, and the worst infestation of locusts in 25 years, which posed a threat to food security. 13 13 See https://www.aljazeera.com/gallery/2020/10/21/in-picture-ethiopia-struggles-to-suppress-desert-locust-infestat 12 Table 5: Percentage of businesses in non-exempt sectors, by turnover quartile Period 1 Period 2 Quartile (turnover) Male-owned Female-owned Male-owned Female-owned 1 26.0 21.4 37.8 2.4 2 19.3 18.3 28.5 4.2 3 24.1 35.1 33.6 12.0 4 46.5 35.5 40.3 27.0 Number of enterprises 476 452 496 445 Source: Author’s calculations using ESS 2018/2019 and 2021/2022. Notes: Significant differences at p < 0.05 are shown in bold. Non-exempt sectors are defined as sectors without a tax-free threshold, for example, bars, cafes, hotels, construction, transit services, and trading in specific products, such as flour, chat, and metals (Council of Ministers Regulation No. 410/2017). Businesses in some activities, particularly in the retail sector, that lack sufficient detail to impute presumptive tax are excluded. Businesses are grouped into turnover quartiles. 4. Empirical methodology The effective tax rate of the presumptive tax is estimated by dividing the tax liability by the turnover, as shown in equation 1. In the literature on corporate tax burdens, this indicator takes a backward-looking approach, which measures the tax burden the enterprises face after all investment and planning decisions have taken place (Nicodème 2002). 14 This is a first-order approximation of the tax burden, which does not account for behavioral responses to taxes. The disadvantage of using turnover (rather than profits) is that it assumes that the profit margins are constant across firms (Nicodème 2002). Nevertheless, turnover is used in this paper because it is the basis for assessing the presumptive tax. Since businesses cannot deduct costs to reduce taxable income in the presumptive tax regime, turnover is equivalent to taxable income. It is also difficult to get reliable estimates of profits in the data because household business expenses cannot be easily separated from the overall household expenses. For example, respondents can misclassify goods and materials for household use as business expenses; owners’ payment to themselves as employee wage costs; and household rent as business rent (see de Mel, McKenzie, and Woodruff, 2009 for a discussion). Effective tax rate = (1) Two measures of tax liabilities are used to estimate the tax burden of the presumptive tax. The first measure is the self-reported tax payments made in the 12 months prior in the ESS data, which is what the businesses claimed to have paid in taxes on income and profits. The second measure is imputed tax 14 In contrast, forward-looking tax rates are calculated to estimate firms’ decisions about investments, financing, or locations (Lazar 2014). 13 liabilities, which is estimated from the government’s presumptive tax code and ESS data on business annual turnover and activity type. The effective tax rates using self-reported data represent the tax burden based on what businesses claim to have paid, while the imputed effective tax rates represent what the tax burden would be using the tax code. The self-reported tax payment in the ESS data measures the taxes paid on income and profits. It is possible that the information provided by respondents could include the turnover tax, which is different from the presumptive tax that businesses pay instead of the VAT as discussed in section 2. This means that the self-reported tax payment could overestimate the presumptive tax payment if it also includes the turnover tax. The analysis compares the self-reported tax payments on income and profits with the imputed presumptive tax liabilities using the tax code, keeping in mind that the self-reported payments may overestimate the presumptive tax because it could include payments for the turnover tax. 15 The imputation method makes the following assumptions. First, it assumes that there is 100% tax compliance. Second, because there are 99 different business activity types in the tax schedule, it is not always clear which tax rate should be applied for tax assessment. For example, since there is no specific rate for electricians, it is not clear whether they should be taxed as “electrical lift works and repair” or as “building contractors”, which has a different tax rate. There is also no clear business activity in the tax schedule for individuals who brew and sell local alcohol, so it is assumed that they are taxed as those in the general merchandise, trade, and grocery activity, which has an exemption threshold for incomes below 50,000 birr. The difficulty in determining the tax rates for each activity highlights how the Revenue Authority tax assessment teams could easily apply different rates for firms engaging in similar activities. Third, detailed information on business activities is required to impute the presumptive tax, but some businesses in the trade sector do not provide sufficient detail about which commodity it was trading in. For example, the tax rates are different for trading in cotton or chat versus trading in vegetables (see Appendix 1). This resulted in some trade enterprises which lacked information on the commodities they were trading in with missing values for the imputed tax liabilities. 5. Results a. Effective tax rates by turnover quartile and sex of business owner First to be estimated is the effective tax rate by turnover quartile using the self-reported method (panel 1) and imputed taxes (panel 2) in Figure 2. In the self-reported method (panel 1), the tax burden falls more 15 The reasons for excluding the turnover tax in the estimation are because this paper focuses on the presumptive tax and there is also not enough information in the ESS data to distinguish locally sold goods and services provided by contractors, grain mills, and combine harvesters (which is charged 2% of turnover) and other services (which is charged 10% of turnover). 14 heavily on the middle quartiles than on the lowest and fourth quartiles. In contrast, if businesses are taxed according to the tax code (panel 2), the presumptive tax is regressive. There is a disproportionate burden on the smallest businesses (in the lowest quartile) who pay 4.3% of turnover in taxes compared to larger businesses (in the highest quartile), who pay 1.5% of turnover using the most recent data. The high effective tax rates that the smallest businesses face using the tax code is particularly concerning because firms could be discouraged from registering and complying with tax payments. The comparison of the effective tax rates using two measures shows that businesses in the highest quartile should be paying more in taxes as the tax rates are higher in the imputed method than in the self- reported method. There could be several reasons for the discrepancy between the two measures. First, some low-revenue businesses operate in sectors without a tax-free threshold and are not paying taxes even though they should (see Appendix 8 for the share of businesses paying taxes by quartile). The lack of an exemption threshold for subsistence businesses explains the regressive nature of the imputed method compared to the self-reported method. Second, there could be a difference in the methods used to estimate annual turnover between the government and this paper. The government assessment of annual revenue is extrapolated from daily revenue with an assumed number of working days whereas this paper estimates annual revenue by multiplying self-reported average monthly income by the number of months of operation. The former method – using daily revenues to extrapolate annual income – is less likely to take seasonal fluctuations in income into account and has been deemed by taxpayers to overestimate turnover and tax liabilities as discussed before. Third, the self-reported tax payment could overestimate the presumptive tax payment if it also includes the turnover tax, which is separate from the presumptive tax. When the results are disaggregated by sex of the owner (panels 3 - 6), the effective tax rates for male businesses in the first quartile are greater than for their female counterparts because a larger share of low revenue male businesses operate in sectors without a tax-free threshold (see Table 5 above). Between the two survey periods, the share of male businesses without a tax-free threshold increased in period 2, while it decreased for female businesses (as discussed in section 3.2), resulting in a larger gender gap in tax burdens for the first quartile in period 2. The effective tax rate for low revenue male businesses using the tax code is particularly high at 12.2%, compared to their female counterparts at 0.4% using the most recent data (panel 6). 15 Figure 2: Effective tax rates (percentage of turnover), by turnover quartile Panel 1: Self-reported Panel 2: Imputed 14.0 14.0 12.0 12.0 10.0 10.0 8.0 8.0 6.5 6.0 6.0 4.3 4.0 1.9 4.0 2.0 1.2 1.4 1.3 0.8 0.6 1.6 1.3 0.6 0.3 2.0 0.7 0.5 1.2 1.5 0.0 1 2 3 4 0.0 1 2 3 4 Turnover quartile Turnover quartile Period 1 Period 2 Period 1 Period 2 Panel 3: Self-reported, period 1 Panel 4: Self-reported, period 2 14.0 14.0 12.0 12.0 10.0 10.0 8.0 8.0 6.0 6.0 4.0 4.0 2.1 2.0 1.2 1.3 1.1 1.6 1.4 1.3 2.0 1.0 2.0 0.8 0.8 0.7 0.3 0.2 0.3 0.0 0.0 0.0 1 2 3 4** 1 2 3 4*** Turnover quartile Tunover quartile Male-owned Female-owned Male-owned Female-owned Panel 5: Imputed, period 1 Panel 6: Imputed, period 2 14.0 14.0 12.2 12.0 12.0 10.0 10.0 7.8 8.0 8.0 5.5 6.0 6.0 4.0 4.0 2.2 1.9 1.3 1.0 1.6 1.3 1.2 2.0 0.5 1.0 2.0 0.4 0.5 0.7 0.2 0.0 0.0 1 2 3* 4 1** 2** 3** 4 Turnover quartile Turnover quartile Male-owned Female-owned Male-owned Female-owned Source: Author’s calculations using ESS 2018/2019 and 2021/2022. Notes: Panels show the effective tax rates (tax as a percentage of turnover) by quartile of turnover. The test of means is significant at *** p<0.01, **p<0.05. 16 b. Effective tax rates by sector and sex of owner Next, we examine the effective tax rates by sex of the business owner, which are presented in rows 1 - 4 in Table 6. The gender difference in tax burden is significant using the tax code with the most recent data (row 2) where male-owned businesses bear a larger tax burden (at 3.1%) than female-owned businesses (at 0.5%). Using self-reported taxes and conditional on positive tax payments (row 3), the effective tax rates are very high for both female businesses and male businesses at 12.4% and 5.0%, respectively, in period 2. Such high rates could be detrimental for business growth and could potentially push them into the informal sector. For example, as a comparison in Ukraine, taxpayers deemed a presumptive tax rate of 6 percent on turnover to be too high, causing discontent in this tax regime (Coolidge and Yilmaz 2016). The results by sector show that the tax burden is low for: (1) the retail trade sector (in rows 7 and 11), as some low revenue firms in this sector (e.g., trading in fruits and vegetables) are exempt, and (2) enterprises that manufacture or sell local beverages (rows 5 and 9) because most fall below the exemption threshold. However, for food and beverage services – a female-dominated sub-sector, female-businesses’ effective tax rates are high (at 13.9% in period 1 and 5.1% in period 2 using the tax code in row 10). Given that low revenue firms are not exempt from the presumptive tax in this sector, the effective tax rate is particularly high. Similarly, low revenue transit services - a male-dominated sector - are not exempt from tax, and so the effective tax rates for male businesses in the transit service sector are also high (at 16.4% in period 1 and 10.3% in period 2 in row 12). The effective tax rates for these two sectors are much higher than for the retail trade sector, which are only 0.9% for male and 0.1% for female businesses with the most recent data. Taken together, the results show that the tax burden of the presumptive tax tends to fall more heavily on the middle quartiles using the self-reported method, but the tax burden using the tax code is regressive. The effective tax rates for male businesses in the lowest quartile are high under the imputed method because 38% operate in sectors without a tax-free threshold. The effective tax rates are also high for the female- dominated food and beverage service sector and for businesses in the transit service, which is male- dominated, because there is no tax exemption for low revenue firms for these sectors. Table 6: Effective tax rates (percentage of turnover), by sector and sex of owner (1) (2) (3) (4) Number of enterprises Period 1 Period 2 Period 1 Period 2 Male- Female- Male- Female- Male- Female- Male- Female- owned owned owned owned owned owned owned owned (1) Effective tax rate (self-reported) 1.4 0.8 0.9 0.9 775 740 703 658 (6.0) (4.6) (3.4) (5.0) 17 (2) Effective tax rate (imputed) 2.5 2.9 3.1 0.5 476 452 496 445 (9.6) (9.9) (9.7) (2.1) (3) Effective tax rate (self- 9.1 10.7 5.0 12.4 161 79 167 84 reported), given tax>0 (12.7) (13.3) (6.6) (14.7) (4) Effective tax rate (imputed), 5.6 10.2 5.2 3.0 276 182 327 166 given tax>0 (13.9) (16.2) (12.1) (4.3) Effective tax rate (self-reported), by sector (5) Manufacturing or sale of local beverages na 0.1 na 0.0 113 94 (0.6) (0.2) (6) Food and beverage services na 0.6 na 1.5 111 61 (2.7) (5.9) (7) Retail trade 0.8 1.1 1.0 1.8 353 389 302 387 (4.1) (5.3) (3.8) (7.2) (8) Transit services 4.1 Na 2.1 na 81 108 (8.0) (4.9) Effective tax rate (imputed), by sector (9) Manufacturing or sale of local na 0.0 na 0.0 113 94 beverages (0.1) (0.1) (10) Food and beverage services na 13.9 na 5.1 100 60 (18.5) (5.6) (11) Retail trade 0.5 0.1 0.9 0.1 139 144 198 214 (2.2) (0.4) (3.2) (0.4) (12) Transit services 16.4 na 10.3 na 80 106 (23.5) (14.8) Source: Author’s calculations using ESS 2018/2019 and 2021/2022. Notes: Imputed presumptive taxes are calculated using the estimated annual turnover and the presumptive tax schedule in the Council of Ministers Regulation No. 410/2017 according to the enterprise activity. This resulted in 599 enterprises with missing values because the activities are not specific enough to assign the presumptive tax values. Significant differences at p < 0.05 are shown in bold. Na indicates sample size is below 50. Numbers in parentheses indicate standard deviation. 6. Policy simulation: A hypothetical presumptive tax with an exemption threshold for microbusinesses and a single tax rate on turnover Given the regressivity of the presumptive tax, a policy is simulated where the tax burden of a hypothetical tax is estimated with an exemption threshold for all firms earning less than 50,000 birr (which is already in place for some sectors), and a single rate on turnover in tax liabilities for firms earning above this threshold. The value of 2.5% on turnover is chosen in this paper as the hypothetical presumptive tax rate for several reasons. There are tradeoffs to consider in setting the tax rate. It should not be set too high such that it discourages formalization, but it should not set be too low so that it could hinder the migration of a growing business from the presumptive tax system to the standard business tax system (Engelschalk and Loeprick 2016). If the rate is set much lower than the standard tax regime, it could encourage businesses 18 to bunch below the standard business tax rate threshold. Yet, it is not easy to compare the effective tax rates between the presumptive tax and the standard business tax because in the presumptive tax system, business expenses, such as labor costs, materials, and depreciation of assets, cannot be deducted from the taxable income when it is assessed on turnover. 16 In the ESS data, profits could not reliably be estimated because there is fungibility between business expenses and household expenses, which makes the effective tax rates comparison between the two tax systems difficult. There are also compliance costs related to paying the standard business taxes because of the requirement for financial record-keeping and the possible need for hiring accountants to comply with taxes. The tax rate of 2.5% is also the percentage that the government could charge without losing tax revenue, which is discussed in more detail in Appendix 9. Table 7 shows the estimated effective tax rates of this hypothetical system and Figure 3 disaggregates the results by turnover quartiles. There are three main findings from the analysis. First, the tax burden is lower for female than male businesses (row 1) because most female enterprises earn less than the 50,000 birr threshold. Second, the effective tax rates for all businesses (row 1) as well as for those conditional on positive tax payments (row 2 in Table 7) under the hypothetical presumptive tax are lower compared to the effective tax rates in the current system (rows 2 and 4 in Table 6, respectively). This is largely due to the exemption of low revenue firms, which reduces the overall effective tax rates. Third, the hypothetical presumptive tax is more progressive, as larger firms would face higher burdens than smaller firms (Figure 3) and larger firms would pay more taxes in the hypothetical system than under the existing system. The results are similar when a 2% tax on turnover is used, except that the effective tax rates are 2% for firms with positive tax liabilities (see Appendix 10). Table 7: Policy Simulation, Effective tax rates of a hypothetical presumptive tax (2.5% of turnover and exemption for businesses earning less than 50,000 birr), by sex of owner Period 1 Period 2 Male-owned Female-owned Male-owned Female-owned (1) Effective tax rate (all) 0.6 0.2 0.9 0.3 (2) Effective tax rate, given tax>0 2.5 2.5 2.5 2.5 Number of enterprises 775 740 703 658 Source: Author’s calculations using ESS 2018/2019 and 2021/2022. Notes: The presumptive taxes are calculated based on 2.5% of turnover and an exemption threshold for firms with a revenue of 50,000 birr or less. Significant differences at p < 0.05 are shown in bold. 16 The business income tax system here refers to Category B taxpayers (unincorporated businesses earning between 500,000 and 1 million birr) and Category A taxpayers (incorporated businesses regardless of turnover or unincorporated businesses earning more than 1 million birr). Categories A and B taxpayers can deduct business expenses from the taxable income whereas Category C taxpayers in the presumptive tax system cannot. 19 Figure 3: Effective tax rates of a hypothetical presumptive tax (2.5% of turnover and exemption for businesses earning less than 50,000 birr), by turnover quartile Panel 1: hypothetical, period 1 Panel 2: hypothetical, period 2 14.0 14.0 12.0 12.0 10.0 10.0 8.0 8.0 6.0 6.0 4.0 4.0 2.5 2.5 1.8 1.4 2.0 2.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.09 0.02 0.0 0.0 1 2 3 4 ** 1 2 3*** 4 Turnover quartile Turnover quartile Male-owned Female-owned Male-owned Female-owned Source: Author’s calculations using ESS 2018/2019 and 2021/2022. Notes: The figure shows the effective tax rates (tax as a percentage of turnover) by quartile of turnover for the hypothetical presumptive tax. The exemption threshold of 50,000 birr is placed within the fourth quartile so some businesses would be exempt even in this quartile. The difference in means between male- and female-owned enterprises is significant at ** p<0.05, *** <0.01. The hypothetical presumptive tax, with an exemption threshold and a flat percentage on turnover in taxes, has several advantages over the current system. First, it is less complicated than the current system which determines tax liabilities based on 1,881 activity-bands. A simpler system would be more transparent and easier for tax assessors and taxpayers to understand. Second, it would eliminate the policy notches that exist within each turnover band in the current system where tax liabilities are lump sum payments according to the turnover bands and activity. Third, the exemption for microbusinesses could facilitate the formalization of informal businesses and bring them into the tax system as the businesses grow. Fourth, it would reduce the activity-specific horizontal inequity the current system has (see Appendix 1 for examples of the tax rates). Fifth, the administration costs of tax collection could be lower because the tax authorities can focus on higher revenue firms instead of collecting taxes door-to-door from micro-subsistence businesses. The possible disadvantages of the hypothetical tax system include: 1) it assumes a fixed profit margin for all firms regardless of sector and profitability; 2) if women operate businesses with lower profit margins than men, a flat tax rate could disproportionately affect women (among those who earn above the tax-free threshold); and 3) with a tax increase for higher revenue businesses, some may choose to stay just below the exemption threshold to avoid paying tax, although it is not clear whether this would occur in Ethiopia. For example, there is lack of evidence of bunching of formal incorporated firms around the VAT threshold 20 in Ethiopia (Mascagni and Molla 2018). It is unclear if a similar pattern would be observed around the presumptive tax exemption threshold. The design of a tax on small and microenterprises needs to consider the tradeoffs between bringing informal businesses into the tax net and the administrative costs (Keen and Mintz 2004). These considerations are particularly important for microenterprises because of their limited revenue generation potential. While medium-sized firms may benefit from tax registration by increasing their customer base and sales, micro-level businesses may not reap the benefits of formalization (see Benhassine et al. 2018 and McKenzie and Sakho 2010 for examples from Benin and Bolivia, respectively). Tax administrations’ capacity in terms of resources and skills may also be limited to monitor and collect taxes from microenterprises. 7. Conclusion This paper examines the equity and gender implications of the presumptive tax on small and microenterprises. Using nationally representative household data and the tax code, the findings show that the tax burden is borne most heavily by the smallest enterprises, thereby making it regressive. The smallest male-owned businesses face a higher effective tax rate compared to female-owned businesses because a larger share of male-owned businesses operate in sectors without an exemption threshold. The results also show that businesses in the food and beverage service subsector, which are predominantly female-owned, and the transit service subsector, which are predominantly male-owned, face a high tax burden because of the lack of a tax-free threshold for these subsectors. Such high tax rates for small enterprises could discourage entrepreneurs from registering their business for taxation and threaten the business sustainability. The presumptive taxation on small and microenterprises in Ethiopia consists of different tax rates for 99 business sectors with 19 turnover bands (with a total of 1,881 activity-bands). The complexity of the tax schedule and the existing process of estimating tax liabilities using daily revenue to extrapolate annual revenues have led to discontent among taxpayers, which could undermine trust towards tax authorities and the perception of tax fairness. An alternative presumptive tax system that is assessed on a fixed percentage on turnover and exempts low revenue firms would reduce the regressivity and promote equity and simplicity. 21 References Akpan, I., Cascant-Sempere, M. J. (2022). Do tax policies discriminate against female traders? A gender framework to study informal market- places in Nigeria. Poverty Public Policy, 14(3), 287-306. Alibhai, S., Buehren, N., Papineni, S., & Pierotti, R. S. (2017). Crossovers: female entrepreneurs who enter male sectors--evidence from Ethiopia. World Bank Policy Research Working Paper, (8065). Anyidoho, N.A., Gallien, M., Ismail, G., Juergens-Grant, F., Rogan, M. & van den Boogaard, V. (2022). Tight tax net, loose safety net: Taxation and social protection in Accra’s informal sector. WIEGO Working Paper No. 45. Manchester, UK: WIEGO. Bachas, P.J., Brockmeyer, A., Dom, R., Semelet, C. M. (2023). Effective Tax Rates and Firm Size. Policy Research Working Paper; no. WPS 10312; Washington, D.C.: World Bank Group. Benhassine, N., McKenzie, D., Pouliquen, V., & Santini, M. (2018). Does inducing informal firms to formalize make sense? Experimental evidence from Benin. Journal of Public Economics, 157, 1-14. Bezabih, M., S. Holden, and A. Mannberg. (2016). The Role of Land Certification in Reducing Gaps in Productivity between Male- and Female-owned Farms in Rural Ethiopia. Journal of Development Studies, 52 (3), 360–76. Bucci, V. (2020). Presumptive taxation methods: A review of the empirical literature. Journal of Economic Surveys, 34(2), 372-397. Caroll, E. (2011). Taxing Ghana’s informal sector: The experience of women [Occasional Paper No.7. Christian Aid]. Retrieved from: http://www.christianaid.org.uk/resources/policy/occasional_paper_series.aspx Council of Ministers Regulation No. 410/2017. Coolidge, J., & Yilmaz, F. (2016). Small Business Tax Regimes What Surveys Reveal about Tax System Use and Abuse. Viewpoint: Public Policy for the Private Sector. Washington, DC: World Bank. © World Bank. de Mel, S., McKenzie, D. and Woodruff, C. (2009). Measuring microenterprise profits: Must we ask how the sausage is made? Journal of Development Economics, 88, 19–31. Dixon, H., Freedman, J. and Yesegat, W. (2019) Small and Micro businesses: Case Studies on the Complexity of ‘Simplified’ Schemes, in Evans, C., Franzsen, R and Stack, E, ed. “Tax Simplification: An African Perspective”, 237-, Pretoria University Law Press, South Africa. Dom, R. and Prichard, W. (2022). Taxing SMEs in Innovations in Tax Compliance. In “Dom, R., Custers, A., Davenport, S.R., Prichard, W. Innovations in Tax Compliance: Building Trust, Navigating Politics, and Tailoring Reform. Washington, DC: World Bank. https://openknowledge.worldbank.org/handle/10986/36946 License: CC BY 3.0 IGO.” Dube, G. and Casale, D. (2019). Informal Sector Taxes and Equity: Evidence from Presumptive Taxation in Zimbabwe. Development Policy Review, Vol. 37, Issue 1, pp. 47-66. Available at SSRN: https://ssrn.com/abstract=3296431 or http://dx.doi.org/10.1111/dpr.12316 Engelschalk, M., & Loeprick, J. (2016). The taxation of micro and small businesses in transition economies: Country experience of the introduction of special tax regimes. Journal of Tax Administration, 2(1), 145-197. Fuje, H. (2017). How Do Women Fare in Rural Nonfarm Economy? Policy Research Working Paper; No. 8080. World Bank, Washington, DC. © World Bank. https://openknowledge.worldbank.org/handle/10986/26952 License: CC BY 3.0 IGO Getachew, A. (2019). Turnover-Based Presumptive Taxation and Taxpayers' Perceptions in Ethiopia. ICTD Working Papers 14551. Institute of Development Studies, International Centre for Tax and Development. Harris, T., and Seid, E. (2021). 2019/20 Survey of the Ethiopian Tax System. IFS Report R187. Institute for Fiscal Studies. https://ifs.org.uk/publications/15402 22 Joshi, A., Prichard, W., & Heady, C. (2014). Taxing the informal economy: The current state of knowledge and agendas for future research. The Journal of Development Studies, 50(10), 1325-1347. DOI: 10.1080/00220388.2014.940910 IMF (2015). Current Challenges in Improving Tax Compliance. Washington, D.C., IMF. https://www.imf.org/en/Publications/Policy-Papers/Issues/2016/12/31/Current-Challenges-in- Revenue-Mobilization-Improving-Tax-Compliance-PP4944. Keen, M., & Mintz, J. (2004). The optimal threshold for a value-added tax. Journal of Public Economics, 88(3-4), 559-576. Komatsu, H., Ambel, A. A., Koolwal, G., & Yonis, M. B. (2022). Gender norms, landholdership, and rural land use fee and agricultural income tax in Ethiopia. Land Use Policy, 121, 106305. Lazăr, S. (2014) Determinants of the Variability of Corporate Effective Tax Rates: Evidence from Romanian Listed Companies, Emerging Markets Finance and Trade, 50:sup4, 113-131, DOI: 10.2753/REE1540-496X5004S4007. Ligomeka, W. (2019). Expensive to be a Female Trader: The Reality of Taxation of Flea Market Traders in Zimbabwe. Working Paper No. 93. International Centre for Tax and Development. Loeprick, J. (2009) Small Business Taxation: Reform to Encourage Formality and Firm Growth, Investment Climate in Practice: Business Taxation No. 1, Washington DC: World Bank. Loening, J., Rijkers, B., Söderbom, M. (2008). Nonfarm Microenterprise Performance and the Investment Climate: Evidence from Rural Ethiopia. Policy Research Working Paper; No. 4577. World Bank, Washington, DC. Marchese, M. (2021). Preferential Tax Regimes for MSMEs: Operational Aspects, Impact Evidence and Policy Implications. ILO Working Paper. https://www.ilo.org/global/publications/working- papers/WCMS_803925/lang--en/index.htm. Mascagni, G., and Molla, K.G. (2018). Business Taxation in Ethiopia: Exploring Administrative Data to Understand Taxpayers’ Responses to Tax Thresholds. IFS Report R189. Institute for Fiscal Studies. Mas-Montserrat, M., Colin, C., Ribault E., and Brys, B. (2023), "The design of presumptive tax regimes", OECD Taxation Working Papers, No. 59, OECD Publishing, Paris. McKenzie, D., & Sakho, Y. S. (2010). Does it pay firms to register for taxes? The impact of formality on firm profitability. Journal of Development Economics, 91(1), 15-24. Meagher, K. (2018) Taxing Times: Taxation, Divided Societies and the Informal Economy in Northern Nigeria, The Journal of Development Studies, 54:1, 1-17. Nicodème, G. (2002). Sector and size effects on effective corporate taxation, Munich Personal RePEc Archive Paper No. 15781. Ohnsorge, F., & Yu, S. eds. (2021). The Long Shadow of Informality: Challenges and Policies. World Bank, Washington, D.C. Paler, L., W. Prichard, R. Sanchez de la Sierra, and C. Samii. (2017). Survey on Total Tax Burden in the DRC, Final Report. Study for the Department for International Development (UK), Kinshasa, Democratic Republic of Congo. Pimhidzai, O., & Fox, L. (2012). Taking from the poor or local economic development: The dilemma of taxation of small informal enterprises in Uganda. CSAE Working Paper. Washington, DC: World Bank. Prichard, W. & van den Boogaard, V. (2017). Norms, Power, and the Socially Embedded Realities of Market Taxation in Northern Ghana. African Studies Review, 60(1): 171-194. Proclamation 410/2017. Federal Income Tax Regulation. Rijkers, B., & Costa, R. (2012). Gender and rural nonfarm entrepreneurship. World Development, 40(12), 2411-2426. Rocha, R., G. Ulyssea, and R. Rachter. (2018). Do Lower Taxes Reduce Informality? Evidence from Brazil. Journal of Development Economics 134 (September): 28-49. Semboja, H.H.H. (2015). Presumptive Tax System and its Influence on the Ways Informal Entrepreneurs Behave in Tanzania. Tanzanian Economic Review, Vol. 5 Nos. 1 & 2, 2015: 72–90. 23 Van den Boogaard, V. (2018). Gender and the formal and informal systems of local public finance in Sierra Leone. Politique africaine, 151(3), 61-86. Van den Boogaard, V., Prichard, W., Jibao, S. (2019). Informal taxation in Sierra Leone: Magnitudes, perceptions and implications. African Affairs, 118(471), 259-284. Waiswa, R., J. Lastunen, G. Wright, M. Noble, J. Okello Ayo, M.Isingoma Nalukwago, T. Kaidu Barugahara, S. Kavuma, I. Arinaitwe, M. Mwesigye, W. Asiimwe and P. Rattenhuber (2021). An Assessment Of Presumptive Tax In Uganda: Evaluating The 2020 Reform And Four Alternative Reform Scenarios Using Ugamod, A Tax-Benefit Microsimulation Model For Uganda. WIDER Working Paper 2021/163. Helsinki: UNU-WIDER. Waseem, M. (2018). Taxes, informality and income shifting: Evidence from a recent Pakistani tax reform. Journal of Public Economics. vol. 157(C), pages 41-77. Workneh, A.M., Baileyegn, E.M., & Stewart-Wilson, G. (2019). Where the Gap Lay: Presumptive Income Tax Assessment for Small and Microenterprises in Addis Ababa City Administration. ICTD Working Paper 94. Woldehanna, T. (2000). Farm and Off-farm Employment in the Tigray Region, Northern Ethiopia: An Economic Analysis and Policy Implications. Ph.D. Thesis. Wageningen University, Netherlands. World Bank (2009). Ethiopia Diversifying the Rural Economy: An Assessment of the Investment Climate for Small and Informal Enterprises. Report No. 49564-ET. World Bank (2016). Tax Compliance Cost Burden and Tax Perceptions Survey in Ethiopia. World Bank, Washington, D.C. https://openknowledge.worldbank.org/handle/10986/24792 License: CC BY 3.0 IGO. World Bank (2018). Poverty and Shared Prosperity 2018: Piecing Together the Poverty Puzzle. World Bank, Washington, D.C. World Bank (2021). The Taxation of Individuals, Households, and Nonfarm Enterprises in Ethiopia: Findings from the 2018/19 Ethiopia Socioeconomic Survey. World Bank, Washington, DC. World Bank (2022). Breaking Barriers: Female Entrepreneurs Who Cross Over to Male-Dominated Sectors. World Bank, Washington, DC. 24 Appendix 1: Presumptive Tax Schedule for Selected Sectors CATEGORY "C" PRESUMPTIVE TAX Presumptive Tax (Birr) Per Year 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 Av 50001-75000 75001-100000 100001-125000 125001-150000 150001-175000 175001-200000 200001-225000 225001-250000 250001-275000 275001-300000 300001-325000 325001-350000 350001-375000 375001-400000 400001-425000 425001-450000 450001-475000 475001-500000 Annual Profit Rate Up to BUSINESS 5000 SECTORS 0 Charcoal Trade 10 0 30 280 530 780 1030 1290 1665 2040 2415 2790 3165 3540 3915 4370 4870 5370 5870 6370 Tailoring Service Only 10 0 30 280 530 780 1030 1290 1665 2040 2415 2790 3165 3540 3915 4370 4870 5370 5870 6370 Cereals and Pulses Trade 10 0 30 280 530 780 1030 1290 1665 2040 2415 2790 3165 3540 3915 4370 4870 5370 5870 6370 Fruits and Vegetables 13 0 255 580 905 1230 1703 2190 2678 3165 3653 4170 4820 5470 6120 6770 7420 8070 8720 9470 Trade Food Items Supply 15 30 405 780 1155 1665 2228 2790 3353 3915 4620 5370 6120 6870 7620 8370 9158 10095 11033 11970 Chat Trade 16 80 480 880 1290 1890 2490 3090 3690 4370 5170 5970 6770 7570 8370 9220 10220 11220 12220 13220 Hotels, Bars and Snacks 25 530 1155 2040 2978 3915 5120 6370 7620 8870 10408 11970 13533 15095 14813 17000 19188 21375 23563 25750 Transit Service 30 780 1665 2790 3915 5370 6870 8370 10095 11970 13845 15720 16125 18750 21375 24000 26625 29250 31875 34500 Source: Council of Ministers Regulation No. 410/20 25 Appendix 2: Percentage of households with a non-farm enterprise (NFE), by sex of household head Period 1 Period 2 2018/2019 2021/2022 % of households with NFEs 22.9 26.8 6770 4959 Period 1 Period 2 Male-headed Female-headed Male-headed Female-headed households households households households % of households with NFEs 22.4 24.4 26.6 27.3 Number of households 4627 2143 3416 1543 % of poor households with NFEs 16.1 19.5 19.8 18.6 % of nonpoor households with NFEs 26.9 27.2 31.4 32.1 Author’s calculations using ESS 2018/2019 (period 1) and 2021/2022 (period 2). Notes: Poor households are defined as those in the bottom 40% of the adult equivalent expenditure distribution. Appendix 3: Average enterprise turnover, by household consumption quartile Panel 1: By household typea Panel 2: By sex of household head Source: Author’s calculations using pooled data from ESS 2018/2019 and 2021/2022. Notes: The graph shows the local polynomial regressions of the average turnover of the enterprise owned by the household (in ‘000 birr, 2019 prices) on the household consumption quartile, where household consumption is the spatially adjusted adult equivalent scale household consumption (in 2019 prices). The sample excludes households that do not own an enterprise. a Female households are households with no male adult present. 26 Appendix 4: Characteristics of enterprise owners Period 1 Period 2 Male owner Female owner Male owner Female owner Age 34.7 33.1 36.4 35.6 Relation to household head (%) Household head 84.7 41.1 73.6 25.3 Spouse of head 2.0 47.2 1.6 55.0 Marital status (%) Married 76.7 57.5 70.1 63.6 Never married, divorced or widowed 23.3 42.5 29.9 36.3 Literacy (%) Can read and write 79.5 51.0 85.8 51.3 Household characteristics % of owners living in female-headed 5.9 46.8 10.5 35.1 household No male adults in household (%) 0.7 31.6 0.0 9.5 Household size 4.6 4.4 5.0 4.8 Dependency ratio 0.84 0.96 0.67 0.71 Number of owners 794 953 724 821 Source: Author’s calculations using ESS 2018/2019 and ESS 2021/2022. Notes: For some enterprises, there is more than one owner. Adults are defined as age 18 or over. There is one male owner of an enterprise, who is 17 years old, living in a household with no male adults. 27 Appendix 5: Primary location where enterprise operates Period 1 (2018/2019) 42.4 % of enterprise 21.5 18.2 18.7 15.6 14.9 15.7 13.5 10.9 9.9 4.8 4.1 4.2 0.0 Inside home*** Outside home** Traditional Shop in Roadside Mobile Construction market commercial location*** site*** area*** Male-owned Female-owned Period 2 (2021/2022) 38.2 % of enterprises 28.9 23.4 17.7 20.1 13.2 13.2 8.4 6.7 8.4 6.4 4.0 3.7 2.2 Inside home*** Outside home Traditional Shop in Roadside Mobile Construction market* commercial location*** site* area** Male-owned Female-owned Note: Author’s calculations using ESS 2018/2019 and 2021/2022. Significant at ***1%, **5%, *10%. 28 Appendix 6: Percentage of households with nonfarm enterprises (restricted to households in both periods 1 and 2) Male or dual adult Female only Percentage of households: household household With an enterprise in period 1 (2018/2019) 22.2 25.2 With an enterprise in period 2 (2021/2022) 27.0 22.6 Reported fewer enterprises in period 2 than in period 1 8.9 9.8 Reported a new enterprise in period 2 15.3 9.6 With a food service enterprise in period 2 1.3 2.3 Reported fewer enterprises in food service in period 2 0.9 3.6 Reported a new food service enterprise in period 2 1.1 1.0 With a transit service enterprise in period 2 1.9 0.0 Reported fewer enterprises in transit service in period 2 0.5 0.1 Reported a new transit service enterprise in period 2 1.7 0.0 Number of households 4,014 835 Source: Author’s calculations using ESS 2018/2019 and 2021/2022. Note: The sample is restricted to all households that are observed in both periods for comparability. It excludes Tigray region. A household is defined to have reported fewer enterprises if the number of enterprises owned in period 2 is less than in period 1. A household reported a new enterprise if the number of enterprises owned in period 2 is greater than in period 1. 29 Appendix 7: Sector distribution of enterprises, by wave (restricted to households in both periods 1 and 2) The figures restrict the sample to households that are observed in both periods 1 and 2 to observe the change in the distribution of sectors between the two periods. The service sector share is lower in period 1 than in period 2 for women and men, but there are gender differences in composition within the service sector. Among female enterprises, the share of food and beverage services – which is a sector without a tax-free threshold – fell from 12.6% in period 1 to 5.5% in period 2, while the share of manufacturing of local beverages increased by 4.5 percentage points. For male enterprises, the “other personal services” category fell from 13.5% to 2.7% and the transportation services increased by 5 percentage points. The proportion of construction also increased among male businesses. Both the transportation services and construction are sectors without a tax-free threshold with implications for the presumptive tax. Period 1, 2018/2019 Trade 46.7 46.2 Manufacture or sale of local beverages 0.4 23.5 Manufacturing clothes 0.3 10.0 Manufacturing food 2.4 6.5 Other manufacturing 2.8 8.6 Food and beverage services and hotels 4.9 12.6 Transportation services 0.1 7.1 Other personal services 6.2 13.5 Construction 1.0 0.0 Others 1.7 5.5 0.0 10.0 20.0 30.0 40.0 50.0 Male-owned Female-owned Period 2, 2021/2022 Trade 44.5 43.1 Manufacture or sale of local beverages 2.6 28.0 Manufacturing clothes 0.9 4.7 Manufacturing food 2.3 7.0 Other manufacturing 3.4 8.1 Food and beverage services and hotels 5.6 5.5 Transportation services 0.1 12.1 Other personal services 1.02.7 Construction 0.3 8.8 Others 5.9 13.4 0.0 10.0 20.0 30.0 40.0 50.0 Male-owned Female-owned Note: Author’s calculations using ESS 2018/2019 and 2021/2022. The sample is restricted to enterprises in households that are included in both periods for comparability. It excludes Tigray region. 30 Appendix 8: Businesses who paid tax, by turnover quartile (%) Period 1 24.3 23.1 16.7 11.7 7.2 8.8 4.6 1.5 1 2 3 4 Turnover quartile Male-owned Female-owned Period 2 29.2 19.4 13.7 13.6 10.1 8.4 4.7 0.4 1 2 3 4*** Turnover quartile Male-owned Female-owned Note: Author’s calculations using ESS 2018/2019 and 2021/2022. ***<0.01. Businesses are grouped into turnover quartiles. Appendix 9: Revenue implications of the Hypothetical Presumptive Tax The hypothetical presumptive tax system in section 6 (with an exemption for low-revenue firms and a 2.5% of turnover in taxes) shows that it is more equitable than the current system. In this Appendix, the possible implications of the hypothetical tax on tax revenue are estimated using the most recent ESS data. In the first scenario, it is assumed that there is a 100% tax compliance rate under the hypothetical presumptive tax system. In the second scenario, it is assumed that only businesses with an annual revenue of over 100,000 birr pay the presumptive tax. The following assumptions are made. First, the compliance rate of higher revenue businesses is assumed to be much greater than the existing system where only 29% of male and 14% of female businesses 31 in the fourth quartile in period 2 pay taxes (see Appendix 8). Second, the self-reported tax payment is assumed to reflect the presumptive tax remitted to government, with the caveat that the self-reported amount might be overestimated because it could include the turnover tax - a separate tax on small businesses. Third, it is assumed that there are no behavioral responses in turnover due to the new tax rates, so the existing turnover is used for estimating the tax revenue in the hypothetical scenarios. Another limitation of the analysis is that the enterprises in the data are representative of those operated by households and are not representative of the business population. For this reason, it is likely that the turnover of the businesses in this data is skewed towards the lower end of the distribution and underestimates the total presumptive tax revenue. Table A9.1 presents the results. Under the assumption of 100% tax compliance (in the first scenario), the share of taxpaying firms at 25% and the estimated revenue (row 2) would be higher than the existing system (row 1). Even if only larger firms with over 100,000 birr pay taxes (in the second scenario), the estimated revenue is higher and about the same proportion of taxpaying firms, suggesting that the hypothetical tax system would likely increase tax revenues if there is tax compliance for higher revenue enterprises (earning over 100,000 birr). In Table A9.2, an alternative hypothetical tax system is shown with a lower presumptive tax rate at 2% of turnover and only firms earning over 100,000 birr remit payments. The estimated tax revenue in the second scenario (compliance by larger firms earning over 100,000 birr) is greater than the current estimated amount, but since the tax compliance is likely to be less than 100%, the results suggest that the tax rate on the turnover should be closer to 2.5% to not lose revenue. 32 Table A9.1: Estimated Tax Revenue (2.5% of turnover and exemption for firms earning 50,000 birr or less), ESS 2021/2022 Percentage of '000,000 taxpaying Existing system birr firms (%) (1) Estimated revenue 1776.3 13.3 First Scenario: 2.5% on turnover,100% tax compliance (2) Estimated revenue under hypothetical system (100% compliance) 4786.9 25.1 Second Scenario: 2.5% on turnover, only tax compliance by larger revenue firms >100,000 birr (3) Estimated revenue under hypothetical system (only by larger revenue firms) 3660.2 13.6 Number of enterprises 1374 Source: Author’s calculations using 2021/2022. a The presumptive taxes under the hypothetical system are calculated based on 2.5% of turnover and an exemption threshold for firms with a revenue of 50,000 birr or less. Appendix A9.2: Estimated Tax Revenue (2% of turnover and exemption for firms earning 50,000 birr or less), ESS 2021/2022 Percentage of '000,000 taxpaying Existing system birr firms (%) (1) Estimated revenue 1776.3 13.3 First Scenario: 2% on turnover, 100% tax compliance (2) Estimated revenue under hypothetical system (100% compliance) 3829.5 25.1 Second Scenario: 2% on turnover, only tax compliance by larger revenue firms >100,000 birr (3) Estimated revenue under hypothetical system (only by larger revenue firms) 2928.1 13.6 Number of enterprises 1374 Notes: Author’s calculations using 2021/2022. a The presumptive taxes under the hypothetical system are calculated based on 2% of turnover and an exemption threshold for firms with a revenue of 50,000 birr or less. Appendix 10: Effective tax rates of a hypothetical presumptive tax, by sex (tax = 2% of turnover) Period 1 Period 2 Male- Female- Male- Female- owned owned owned owned (1) Effective tax rate (all) 0.5 0.2 0.7 0.2 (2) Effective tax rate, given tax>0 2.0 2.0 2.0 2.0 Number of enterprises 775 740 703 658 Source: Author’s calculations using ESS 2018/2019 and 2021/2022. Notes: The presumptive taxes are calculated based on 2% of turnover and an exemption threshold for firms with a revenue of 50,000 birr or less. Significant differences at p < 0.05 are shown in bold. 33